CardioExchange Archive

After the presentation and publication of the pivotal ARISTOTLE trial,  the novel anticoagulant apixaban (Eliquis, Pfizer and Bristol-Myers Squibb) was widely expected to be  a blockbuster. But then it got bogged down at the FDA where initial hopes for a speedy approval were dashed after highly critical reviews. Approval of the drug was delayed for 9 months past the original deadline as a result of both a Prescription Drug User Fee Act (PDUFA) date extension and a complete response letter from the FDA. Now an article by Sue Sutter in  Pharmaceutical Approvals Monthly, based on documents posted by the FDA on its website, offers an inside look at the drug’s roller coaster ride through the FDA.

The article describes in detail the efforts of the FDA to investigate evidence of fraud and dispensing errors in ARISTOTLE. As Sutter writes: “To many looking in from the outside, the apixaban NDA filing seemed like a sure thing.” Initial results of ARISTOTLE suggested that it had “the potential to be best-in-class if the published ARISTOTLE data were to be believed.”

It was not publicly revealed at the time, but the FDA documents now show that the reason for the original 3-month extension of the PDUFA date was serious concerns about trial misconduct in China. According to Sutter, a senior clinical site manager, with another monitor, “altered source records to cover up evidence of good clinical practice violations.” This eventually led FDA investigators to question data from 24 of the 36 sites in China.

Ultimately, the FDA reviewers concluded that the results of the trial were not substantially altered by the problem. Sutter writes: “Thus, while the fraud in China did not derail the apixaban application, it appears to have set the review back by three months and, in the process, highlighted clinical reviewers’ larger concerns about sloppy execution and inadequate sponsor oversight of the pivotal trial.”

A second major problem, and the one that ultimately led to the complete response letter, was that the companies were unable to demonstrate  that patients received their assigned medications. This was a potential source of enormous trouble because of the study’s double-dummy design, creating the possibility that patients might have received either both active drugs, putting them at high risk for bleeding, or two placebos, putting them at high risk for stroke.

“Questions surrounding the extent of dispensing errors in ARISTOTLE not only raised concerns about how such errors affected study outcomes, they also led FDA reviewers to question the quality of monitoring and oversight of the trial,” writes Sutter. One FDA official wrote that if the agency had been aware of these errors prior to companies submitting the NDA, “we would have refused to file it.”

Three months after receiving the complete response letter the companies resubmitted the NDA. The FDA reviewers then concluded that the major findings of the trial were valid, leading to approval of the drug on December 28, 2012.

A second article by Sutter examines the controversy over the wording and placement of the mortality finding of ARISTOTLE in the Eliquis label. Although apixaban was associated with a statistically significant reduction in mortality in ARISTOTLE, the marginal result (p=0.047) was not strong enough for the FDA. Writes Sutter: “While Eliquis’ sponsors surely would have liked an Indications statement for reduced all-cause mortality, FDA declined to give apixaban such a competitive advantage over Pradaxa and Xarelto.” Instead, the mortality “information is relegated to the Adverse Reactions and Clinical Studies section of the label.”

Sutter reports that there are strong hints in the redacted FDA documents that Pfizer and BMS “requested a mortality reduction claim as part of the proposed Indications statement for Eliquis.” According to Sutter:

The NDA requested approval of Eliquis ‘to reduce the risk of stroke, systemic embolism, [redacted] in patients with nonvalvular atrial fibrillation,’ according to various review documents. A second line of the proposed Indications statement also is redacted.”

Sutter quotes the FDA reviewers:

‘The mortality finding (superiority for all-cause death) is not nearly as robust as the findings for the primary endpoint and major bleeding,’ the reviewers said, noting that the agency’s statistical reviewer calculated that one less death in the warfarin arm would negate statistical significance for apixaban’s superiority.”

Asked for a response to the articles, Pfizer and BMS sent the following statement:

Bristol-Myers Squibb and Pfizer are committed to achieving the highest level of scientific rigor and adherence to the principles of good clinical practice in all clinical development programs. In response to the Complete Response Letter from the FDA on Eliquis, the companies conducted significant analyses to confirm that although there were medication dispensing errors in ARISTOTLE, the rate was very low and it did not impact the outcome measures of the trial. The questions in the complete response letter were resolved to the FDA’s satisfaction and Eliquis was approved by the agency on December 28, 2012, ahead of the March 17, 2013, resubmission PDUFA date. Bristol-Myers Squibb and Pfizer believe that the FDA approval of Eliquis to reduce the risk of stroke and systemic embolism in patients with nonvalvular atrial fibrillation reflects the strength of the clinical data for Eliquis in this indication.”

Ethan Weiss, a cardiologist at the University of California San Francisco Medical Center, feels strongly that the mortality issue is a red herring. A strong supporter of apixaban, he is most impressed by the safety profile of the drug. The lower rate of intracerebral hemorrhage — the most feared complication of anticoagulant therapy — associated with apixaban in ARISTOTLE means “it is the clear winner.” He wrote me:

To me, the major benefit of this drug is its safety. I am much less impressed by its efficacy, but honestly a drug that is non-inferior to warfarin is all that any of us wanted for a long time. So the most conservative way to look at it is that we have a warfarin replacement that 1) is at least as good as warfarin in terms of stroke prevention, 2) is demonstrably safer in terms of major bleeds (most notably in the head), 3) requires no monitoring, 4) has many fewer drug-drug or drug-food interactions. Put that together and you have a big win.”